Cancer is the second leading cause of death in the US. The majority of cancers are of epithelial origin. Early diagnosis of pre-invasive epithelial neoplasia can dramatically reduce the incidence and mortality of carcinoma. Thus, there is a desperate need for highly sensitive and cost-effective screening and diagnostic techniques to identify curable precancerous lesions. Epithelial pre-cancers are characterized by a variety of architectural and morphological features including increased nuclear size, increased nuclear/cytoplasmic ratio, hyperchromasia and pleomorphism. In addition, there is increasing evidence of significant changes occurring in the stromal layer at the earliest stages of carcinogenesis as a result of epithelial-stromal interactions. A major limitation of current clinical diagnosis is that morphological and molecular changes associated with early carcinogenesis can be assessed only after invasive biopsy. Optical techniques can assess morphologic and biochemical alterations in epithelial tissue, non-invasively and in real-time. The goat of this proposal is to develop a new technology based on polarized reflectance spectroscopy and imaging to enable non-invasive real-time detection and monitoring of morphological and architectural changes in epithelium and underlying connective tissue associated with carcinogenesis. The aims of this proposal are to: (1) develop theoretical models for polarized reflectance spectroscopy of epithelial tissue and stroma; (2) to test the theoretical predictions in tissue models of human epithelium at different stages of cancer progression; (3) to use the theoretical models to determine the sensitivity of polarized reflectance spectroscopy to morphology of epithelial nuclei and morphology of stroma; (4) to design, construct and test a new fiber optic endoscope for polarized reflectance spectroscopy in vivo; (5) to evaluate the endoscope in clinical studies of oral cavity mucosa. (6) to develop an imaging instrument based on the principles of polarized illumination and detection and to test this device in clinical trials. The inexpensive optical sensors proposed here can immediately benefit health care by reducing the number of unnecessary biopsies, enabling combined diagnosis and therapy, and reducing the need for clinical expertise.